INDEX

Explanations

the possessive form of the word "it"

oai_token-act-pair · gpt-3.5-turbo

instances of the contraction "it's."

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

scope

-0.62

={

-0.61

ESE

-0.61

-0.60

 Deaths

-0.59

rating

-0.59

rette

-0.59

andise

-0.59

umbnail

-0.59

igraph

-0.58

POSITIVE LOGITS

 gonna

1.07

 gotta

1.04

 unclear

0.92

 been

0.89

 worth

0.89

got

0.83

 impossible

0.82

etsk

0.82

 easy

0.80

Ĥİ

0.80

Activations Density 0.080%

the possessive form of the word "it"

instances of the contraction "it's."

No Comments

No Known Activations