INDEX

Explanations

medical and ethical concepts or phrases related to destruction or negativity

oai_token-act-pair · gpt-3.5-turbo

terms related to various forms of crises and challenges

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

[/

-0.61

[/

-0.60

CFR

-0.60

 Galactic

-0.54

annis

-0.54

||

-0.53

 coerc

-0.53

urations

-0.53

stant

-0.52

 aggregate

-0.52

POSITIVE LOGITS

tackle

0.69

 guiActiveUn

0.65

è£ħ

0.64

omas

0.63

abilia

0.61

reet

0.60

themed

0.59

é¾įåĸļå£«

0.58

ibles

0.57

raft

0.56

Activations Density 1.131%

medical and ethical concepts or phrases related to destruction or negativity

terms related to various forms of crises and challenges

No Comments

No Known Activations