INDEX

Explanations

phrases indicating the direction of an action or relationship

oai_token-act-pair · gpt-3.5-turbo

phrases that suggest alternative perspectives or methods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ĸļ

-1.03

uster

-0.91

usters

-0.91

egu

-0.64

 Parenthood

-0.63

 subsequ

-0.63

OU

-0.61

 appre

-0.60

nutrition

-0.58

ITED

-0.58

POSITIVE LOGITS

fare

1.03

finding

0.86

 round

0.78

ward

0.77

trip

0.76

point

0.72

 Around

0.69

points

0.69

kell

0.68

forward

0.68

Activations Density 0.052%

phrases indicating the direction of an action or relationship

phrases that suggest alternative perspectives or methods

No Comments

No Known Activations