INDEX

Explanations

mentions of people's names or specific identifying terms

oai_token-act-pair · gpt-3.5-turbo

specific references to notable individuals and their actions or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

toggle

-0.74

ppo

-0.66

Enlarge

-0.64

»Ĵ

-0.63

rued

-0.62

ilde

-0.61

terday

-0.61

aughs

-0.61

ratulations

-0.60

onnaissance

-0.60

POSITIVE LOGITS

to

0.99

 favorably

0.98

 trustworthy

0.89

 unfit

0.84

 capable

0.84

 possesses

0.84

 incapable

0.84

 guilty

0.82

 unreliable

0.78

 lacks

0.78

Activations Density 0.343%

mentions of people's names or specific identifying terms

specific references to notable individuals and their actions or situations

No Comments

No Known Activations