INDEX

Explanations

number-related terms and statistical information

oai_token-act-pair · gpt-3.5-turbo

instances of significant risks and concerns

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

west

-0.70

 Havana

-0.66

opolis

-0.66

 Camer

-0.65

 Bras

-0.64

apolis

-0.64

gow

-0.64

Virgin

-0.63

arta

-0.62

 Univ

-0.62

POSITIVE LOGITS

 descriptions

0.99

 clauses

0.98

 ones

0.97

 verbs

0.96

 categories

0.94

 acron

0.92

 occurrences

0.92

 adject

0.90

 prohibitions

0.90

 phrases

0.90

Activations Density 0.841%

number-related terms and statistical information

instances of significant risks and concerns

No Comments

No Known Activations