INDEX

Explanations

, based on the activations provided, it seems like this neuron is finding words indicating sequence or ordering, alongside a specific focus on numerical sequences

oai_token-act-pair · gpt-3.5-turbo

instances of punctuation, specifically commas

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 enthus

-0.69

Attach

-0.65

Deal

-0.65

Catalog

-0.64

MpServer

-0.64

Ult

-0.62

adden

-0.62

ForgeModLoader

-0.61

GF

-0.61

Cra

-0.60

POSITIVE LOGITS

 however

0.83

 although

0.78

 namely

0.76

 though

0.75

0.68

 according

0.64

hester

0.64

 moreover

0.64

 Shank

0.63

 please

0.63

Activations Density 0.106%

, based on the activations provided, it seems like this neuron is finding words indicating sequence or ordering, alongside a specific focus on numerical sequences

instances of punctuation, specifically commas

No Comments

No Known Activations