INDEX

Explanations

texts related to various types of discussions, debates, and expert opinions on specific subjects or topics

oai_token-act-pair · gpt-3.5-turbo

phrases related to discussion topics and subjects of debate

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

 Towers

-0.83

iaries

-0.75

lett

-0.72

ann

-0.69

endars

-0.67

ository

-0.65

 juven

-0.65

eele

-0.63

anova

-0.62

addon

-0.62

POSITIVE LOGITS

 itself

0.79

atics

0.74

 solving

0.68

ophys

0.67

 anew

0.66

 proposition

0.66

 peacefully

0.66

osphere

0.63

HRC

0.63

ultimate

0.63

Activations Density 0.152%

texts related to various types of discussions, debates, and expert opinions on specific subjects or topics

phrases related to discussion topics and subjects of debate

No Comments

No Known Activations