INDEX

Explanations

phrases that include the word "points" followed by a numeric value

oai_token-act-pair · gpt-3.5-turbo

phrases relating to scoring or points in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

uthor

-0.81

 delinqu

-0.73

yss

-0.70

ntil

-0.69

ufact

-0.68

anical

-0.66

 notor

-0.66

ÄŁ

-0.64

ctors

-0.63

ourt

-0.63

POSITIVE LOGITS

manship

0.98

points

0.95

 points

0.88

point

0.87

Points

0.78

pread

0.75

afety

0.73

lash

0.73

 point

0.73

pointers

0.72

Activations Density 0.035%

phrases that include the word "points" followed by a numeric value

phrases relating to scoring or points in various contexts

No Comments

No Known Activations