INDEX

Explanations

words related to features and specifications, particularly those that are built-in or pre-installed

oai_token-act-pair · gpt-3.5-turbo

phrases related to competition or comparison

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Â¯

-0.52

 Somerset

-0.48

 orderly

-0.44

jog

-0.44

 Waste

-0.42

 indoors

-0.42

Sit

-0.42

 Blind

-0.42

 dstg

-0.42

 respir

-0.41

POSITIVE LOGITS

ãĤ¤

0.61

ĵ

0.61

ographies

0.60

ī

0.59

ãĥł

0.58

åĬ

0.57

ages

0.57

izations

0.57

ĩ

0.56

ups

0.56

Activations Density 0.554%

words related to features and specifications, particularly those that are built-in or pre-installed

phrases related to competition or comparison

No Comments

No Known Activations