INDEX

Explanations

specific notations or keywords, such as 'Ċ' and numbers that seem to be categorized

oai_token-act-pair · gpt-3.5-turbo

specific patterns or structures in enumerated lists or bulleted points

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

abouts

-0.60

 accordingly

-0.59

hement

-0.59

heet

-0.57

 heading

-0.57

Mad

-0.56

 thrust

-0.56

 winding

-0.55

cence

-0.54

 proud

-0.54

POSITIVE LOGITS

âĹı

1.11

Â·

0.94

ÂłÂłÂłÂł

0.90

ÂłÂłÂł

0.87

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

0.79

ÂłÂłÂłÂłÂłÂłÂłÂł

0.79

pmwiki

0.79

âĸł

0.79

Pear

0.79

âĺħ

0.79

Activations Density 0.105%

specific notations or keywords, such as 'Ċ' and numbers that seem to be categorized

specific patterns or structures in enumerated lists or bulleted points

No Comments

No Known Activations