INDEX

Explanations

be aware of references to a stated or presumed truth

oai_token-act-pair · gpt-3.5-turbo

mentions of the concept of "fact"

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

avorite

-0.81

 livest

-0.75

 throats

-0.73

 throat

-0.70

airs

-0.67

itsch

-0.66

interstitial

-0.65

 lungs

-0.65

ESE

-0.60

incinn

-0.60

POSITIVE LOGITS

ually

1.14

orial

1.10

uality

1.07

ional

0.99

itious

0.98

oids

0.95

uate

0.87

uated

0.83

finding

0.82

icity

0.81

Activations Density 0.024%

be aware of references to a stated or presumed truth

mentions of the concept of "fact"

No Comments

No Known Activations