INDEX

Explanations

the word "Du" followed by numbers

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Du" in various contexts

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

ãģĨ

-0.75

 Wanted

-0.73

mberg

-0.73

giving

-0.72

ISC

-0.71

nikov

-0.71

 Locked

-0.68

 Across

-0.68

ħĭ

-0.67

ãĤ¤ãĥĪ

-0.67

POSITIVE LOGITS

pee

0.95

pling

0.93

Du

0.90

ente

0.87

isine

0.86

pees

0.85

cci

0.85

Pont

0.84

ples

0.83

alog

0.82

Activations Density 0.005%

the word "Du" followed by numbers

mentions of the name "Du" in various contexts

No Comments

No Known Activations