INDEX

Explanations

occurrences of a specific word

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific surname or term related to individuals involved in a news context

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

weeney

-0.77

xon

-0.72

urgical

-0.63

asers

-0.63

xual

-0.63

 Caldwell

-0.61

bard

-0.60

 Cobra

-0.59

appropriate

-0.59

SEC

-0.58

POSITIVE LOGITS

itives

1.17

erer

0.78

sworth

0.75

istani

0.73

aby

0.73

ief

0.73

ened

0.72

itive

0.72

ries

0.70

Activations Density 0.013%

occurrences of a specific word

mentions of a specific surname or term related to individuals involved in a news context

No Comments

No Known Activations