INDEX

Explanations

adjectives that describe something intensively or extremely

oai_token-act-pair · gpt-3.5-turbo

words that indicate action or conditions associated with processes or changes

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 settlement

-0.73

CRIPTION

-0.69

 throats

-0.69

SHIP

-0.69

 deduction

-0.65

 snippet

-0.64

 divergence

-0.64

 lapse

-0.64

 designation

-0.62

 transitions

-0.62

POSITIVE LOGITS

umbered

1.10

icent

1.10

ussed

1.08

actly

1.08

alled

1.06

rolled

1.05

icked

1.02

inged

1.02

usted

1.01

iably

1.00

Activations Density 0.252%

adjectives that describe something intensively or extremely

words that indicate action or conditions associated with processes or changes

No Comments

No Known Activations