INDEX

Explanations

proper names and titles of individuals in an article

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

proper nouns, especially names of individuals involved in discussions or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

âĢ¦âĢ¦âĢ¦âĢ¦

-0.73

 almost

-0.73

 compromise

-0.73

 catch

-0.73

 dealing

-0.72

 quite

-0.71

 life

-0.71

 satisfied

-0.71

 exactly

-0.70

 ever

-0.70

POSITIVE LOGITS

ilyn

1.02

idth

1.00

ricia

0.91

antha

0.88

 Eisen

0.85

 Beir

0.85

arie

0.85

orio

0.84

oslav

0.84

veyard

0.84

Activations Density 0.511%

proper names and titles of individuals in an article

proper nouns, especially names of individuals involved in discussions or events

No Comments

No Known Activations