INDEX

Explanations

phrases related to specific locations and events

oai_token-act-pair · gpt-3.5-turbo

references to specific events or significant occurrences

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

NetMessage

-0.96

 Flavoring

-0.73

limits

-0.66

Ħ¢

-0.63

criminal

-0.57

advertising

-0.55

 detriment

-0.54

 compromising

-0.54

 undermining

-0.53

]."

-0.53

POSITIVE LOGITS

iversary

0.79

 afterwards

0.77

sequently

0.75

 afterward

0.72

 reception

0.70

 later

0.70

 thereafter

0.68

 followed

0.68

later

0.66

 subsequently

0.66

Activations Density 1.129%

phrases related to specific locations and events

references to specific events or significant occurrences

No Comments

No Known Activations