INDEX

Explanations

words related to discussion moderation or influence

oai_token-act-pair · gpt-3.5-turbo

terms related to data evaluation and categorization

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

istar

-0.87

TIT

-0.67

 Enhancement

-0.66

 interrogation

-0.66

heid

-0.65

ortium

-0.65

ctrl

-0.63

illin

-0.63

mingham

-0.62

LOD

-0.62

POSITIVE LOGITS

ske

1.15

 skew

0.99

wn

0.86

aling

0.83

elling

0.81

lling

0.79

eling

0.78

ldom

0.78

eled

0.78

ling

0.77

Activations Density 0.017%

words related to discussion moderation or influence

terms related to data evaluation and categorization

No Comments

No Known Activations