INDEX

Explanations

phrases introducing information or attributing statements to sources

oai_token-act-pair · gpt-3.5-turbo

references to sources or authorities in a statement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

apons

-0.74

ãĥĵ

-0.71

 ashore

-0.70

 realise

-0.69

estern

-0.68

bered

-0.67

ãĥĨãĤ£

-0.67

ãģĹ

-0.67

ãĤ¨ãĥ«

-0.65

POSITIVE LOGITS

chwitz

0.94

Format

0.83

Ĥİ

0.82

Style

0.79

format

0.79

ety

0.77

 Sources

0.73

Rank

0.72

graded

0.68

ymm

0.67

Activations Density 0.041%

phrases introducing information or attributing statements to sources

references to sources or authorities in a statement

No Comments

No Known Activations