INDEX

Explanations

commands related to physical actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "press" and related variations

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

 Trayvon

-0.68

 Cloak

-0.67

cknowled

-0.63

anie

-0.62

aez

-0.62

SEE

-0.60

iland

-0.58

 Foss

-0.58

 Emin

-0.57

nown

-0.57

POSITIVE LOGITS

urized

1.32

press

0.86

ured

0.82

 presses

0.82

ures

0.82

ature

0.78

ername

0.77

ur

0.75

 pressed

0.73

aic

0.72

Activations Density 0.033%

commands related to physical actions

instances of the word "press" and related variations

No Comments

No Known Activations