INDEX

Explanations

personal pronouns used for self-referencing actions

oai_token-act-pair · gpt-3.5-turbo

phrases that contain the word "themselves."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

grade

-0.75

amia

-0.73

order

-0.70

execute

-0.69

ulton

-0.68

aster

-0.67

asia

-0.67

ritz

-0.66

pour

-0.66

pak

-0.65

POSITIVE LOGITS

selves

1.18

 tremend

0.98

 selves

0.93

self

0.91

 exting

0.88

 themselves

0.87

 conduc

0.86

 proport

0.84

 exha

0.82

 behavi

0.79

Activations Density 0.027%

personal pronouns used for self-referencing actions

phrases that contain the word "themselves."

No Comments

No Known Activations