INDEX

Explanations

names of famous people or places

oai_token-act-pair · gpt-3.5-turbo

references to the concept of fame

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

THER

-0.74

heed

-0.73

ilation

-0.73

otent

-0.72

adies

-0.72

pt

-0.70

opt

-0.66

©¶æ

-0.66

IDES

-0.66

required

-0.66

POSITIVE LOGITS

rities

0.89

 landmarks

0.86

 famous

0.77

 nickname

0.73

NESS

0.72

 comeback

0.71

 Circus

0.70

 BaseType

0.70

 recre

0.69

 longevity

0.69

Activations Density 0.019%

names of famous people or places

references to the concept of fame

No Comments

No Known Activations