INDEX

Explanations

mentions of a negative event related to law enforcement or politics

oai_token-act-pair · gpt-3.5-turbo

the negation of statements or terms that indicate disagreement

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

 tremend

-0.84

 behavi

-0.78

 rake

-0.76

 hemor

-0.75

 Reloaded

-0.70

 indo

-0.70

 lapt

-0.69

 Palestin

-0.68

 exha

-0.68

 Practices

-0.67

POSITIVE LOGITS

ween

1.11

weet

1.02

itles

1.02

otally

1.01

ople

0.98

apest

0.98

urtle

0.97

ruly

0.95

iful

0.94

ouched

0.93

Activations Density 0.056%

mentions of a negative event related to law enforcement or politics

the negation of statements or terms that indicate disagreement

No Comments

No Known Activations