INDEX

Explanations

references to new or fresh items or entities

oai_token-act-pair · gpt-3.5-turbo

mentions of brands or brand-related terms

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

ulhu

-0.94

pmwiki

-0.87

veyard

-0.80

 Poverty

-0.73

abama

-0.72

SEE

-0.71

cale

-0.69

poons

-0.69

 Able

-0.69

vae

-0.69

POSITIVE LOGITS

ishing

0.98

 loyalty

0.96

enburg

0.87

ished

0.86

brand

0.82

ages

0.79

Brand

0.77

aging

0.77

ishes

0.77

 brand

0.72

Activations Density 0.016%

references to new or fresh items or entities

mentions of brands or brand-related terms

No Comments

No Known Activations