INDEX

Explanations

action-related phrases with a clear purpose

oai_token-act-pair · gpt-3.5-turbo

phrases expressing intentions, aims, hopes, or goals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

avorite

-0.86

 nodd

-0.73

elta

-0.70

imore

-0.64

ocket

-0.64

aukee

-0.59

zek

-0.59

 batted

-0.58

heed

-0.57

oggles

-0.57

POSITIVE LOGITS

UTH

0.74

guyen

0.69

 à¨

0.66

ĺħ

0.66

onym

0.66

 caveat

0.63

ims

0.62

 implication

0.60

 either

0.59

 pretext

0.59

Activations Density 0.097%

action-related phrases with a clear purpose

phrases expressing intentions, aims, hopes, or goals

No Comments

No Known Activations