INDEX

Explanations

phrases related to extremes or intensities, such as "too far" or "extremely hard"

oai_token-act-pair · gpt-3.5-turbo

phrases indicating extremity or excessiveness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ilies

-0.64

ility

-0.62

 restraint

-0.60

ilege

-0.60

ability

-0.59

catentry

-0.57

æĿ

-0.57

ulty

-0.57

 Valid

-0.57

 disposable

-0.55

POSITIVE LOGITS

Ô

0.86

wagon

0.83

iral

0.82

 unnoticed

0.81

eper

0.79

 route

0.79

 overboard

0.79

bye

0.78

nova

0.77

 lengths

0.75

Activations Density 0.153%

phrases related to extremes or intensities, such as "too far" or "extremely hard"

phrases indicating extremity or excessiveness

No Comments

No Known Activations