INDEX

Explanations

proper nouns and organizations related to statements and official releases

oai_token-act-pair · gpt-3.5-turbo

frequent instances of the word 'statement'

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

fights

-0.74

cells

-0.72

 experimented

-0.71

effects

-0.71

 mills

-0.69

anism

-0.67

 intersections

-0.67

 è£ıè

-0.67

 eyed

-0.67

 Units

-0.67

POSITIVE LOGITS

 scathing

1.25

 statement

1.21

 lengthy

1.18

 slew

1.11

 memo

1.09

 letter

1.08

 memorandum

1.05

 manifesto

1.05

 pamphlet

1.03

 brief

1.03

Activations Density 0.208%

proper nouns and organizations related to statements and official releases

frequent instances of the word 'statement'

No Comments

No Known Activations