INDEX

Explanations

sequences of characters that don't form readable English words or phrases

oai_token-act-pair · gpt-3.5-turbo

sequences of uppercase letters or codes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

theless

-0.91

 conduc

-0.80

 confir

-0.80

ailability

-0.79

âĸ¬

-0.75

enhagen

-0.74

 conclud

-0.71

 enthusi

-0.70

 withd

-0.70

 nomine

-0.70

POSITIVE LOGITS

fc

0.81

dn

0.78

_-

0.77

cd

0.75

df

0.74

kw

0.74

wl

0.74

0.72

__

0.72

ax

0.72

Activations Density 0.331%

sequences of characters that don't form readable English words or phrases

sequences of uppercase letters or codes

No Comments

No Known Activations