INDEX

Explanations

words or phrases related to being physically placed or located somewhere on an object or person

oai_token-act-pair · gpt-3.5-turbo

references to labeling or identifying objects

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

imore

-0.66

IAN

-0.65

ACE

-0.64

MpServer

-0.58

 Understanding

-0.56

USS

-0.56

 folly

-0.53

IDA

-0.53

ugh

-0.52

ubi

-0.51

POSITIVE LOGITS

tow

0.96

tones

0.94

 them

0.82

sole

0.79

 forefront

0.74

Interstitial

0.69

abouts

0.69

 front

0.67

yx

0.66

Top

0.64

Activations Density 0.202%

words or phrases related to being physically placed or located somewhere on an object or person

references to labeling or identifying objects

No Comments

No Known Activations