INDEX

Explanations

The neuron is looking for terms related to comparisons and changes in pace

oai_token-act-pair · gpt-3.5-turbo

phrases related to keeping up with progress or maintaining pace

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

rique

-0.67

ells

-0.64

ums

-0.63

ints

-0.62

ell

-0.62

int

-0.61

jun

-0.61

umb

-0.60

kinson

-0.60

ogi

-0.59

POSITIVE LOGITS

Behind

0.75

WARD

0.74

daq

0.73

ciating

0.73

Sharp

0.71

DragonMagazine

0.71

atana

0.70

lished

0.70

ilibrium

0.68

enance

0.67

Activations Density 0.329%

The neuron is looking for terms related to comparisons and changes in pace

phrases related to keeping up with progress or maintaining pace

No Comments

No Known Activations