INDEX

Explanations

verbs related to causing something to become in a specific state or condition

oai_token-act-pair · gpt-3.5-turbo

instances of the word "render" and its derivatives, particularly in contexts related to making something ineffective or altered

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

!/

-0.76

andise

-0.72

Cly

-0.72

swick

-0.67

bies

-0.66

FTWARE

-0.64

 Compton

-0.63

ron

-0.62

gow

-0.60

stre

-0.59

POSITIVE LOGITS

ezvous

1.03

 moot

0.94

 obsolete

0.93

 useless

0.82

 unus

0.82

 irrelevant

0.81

 harmless

0.78

gements

0.78

ings

0.77

 invalid

0.77

Activations Density 0.068%

verbs related to causing something to become in a specific state or condition

instances of the word "render" and its derivatives, particularly in contexts related to making something ineffective or altered

No Comments

No Known Activations