INDEX

Explanations

words related to ranks or titles

oai_token-act-pair · gpt-3.5-turbo

punctuation and stylized text formats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

IDENT

-0.64

aminer

-0.63

Poll

-0.63

IRC

-0.62

 Bomber

-0.62

Alert

-0.60

velt

-0.59

 Platform

-0.58

LECT

-0.58

TPP

-0.57

POSITIVE LOGITS

0.87

0.72

ahs

0.67

acia

0.67

ta

0.65

Ãł

0.65

Ã©s

0.64

ataka

0.64

nen

0.61

dds

0.60

Activations Density 0.051%

words related to ranks or titles

punctuation and stylized text formats

No Comments

No Known Activations