INDEX

Explanations

movie titles

oai_token-act-pair · gpt-3.5-turbo

entities related to entertainment and media, such as films, shows, songs, games, and events

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

stasy

-0.69

ilaterally

-0.68

-0.67

externalActionCode

-0.67

Enlarge

-0.64

âĢ¢âĢ¢

-0.63

ãģĨ

-0.61

 âĢ¦"

-0.60

ceive

-0.59

Semitism

-0.58

POSITIVE LOGITS

 consists

0.93

 consisted

0.92

 also

0.88

 revolves

0.87

 culmin

0.81

 comprises

0.81

 itself

0.79

 however

0.76

 reportedly

0.75

 includes

0.75

Activations Density 0.442%

movie titles

entities related to entertainment and media, such as films, shows, songs, games, and events

No Comments

No Known Activations