INDEX

Explanations

statements involving expert opinions

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate expert opinions or statements

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

 ILCS

-0.89

WithNo

-0.78

 Himself

-0.73

ãĥĺ

-0.71

à¤

-0.68

ONSORED

-0.68

=~=~

-0.66

TAG

-0.65

ãģ£

-0.65

Interstitial

-0.64

POSITIVE LOGITS

 alike

0.78

seq

0.75

 departures

0.73

vae

0.69

orians

0.68

selves

0.68

 anecd

0.66

Rao

0.62

uti

0.61

hran

0.60

Activations Density 0.221%

statements involving expert opinions

phrases that indicate expert opinions or statements

No Comments

No Known Activations