INDEX

Explanations

the name "Leonard" with varying levels of activation

oai_token-act-pair · gpt-3.5-turbo

the name "Leonard" 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

skirts

-0.75

tch

-0.73

manship

-0.73

PLA

-0.72

ħĭ

-0.71

exempt

-0.70

illance

-0.69

pron

-0.69

sites

-0.67

POSITIVE LOGITS

 Leonard

1.04

 Lauder

0.88

 Cohen

0.88

angelo

0.82

 Bernstein

0.78

 Hayward

0.78

ergic

0.76

 Strauss

0.76

antine

0.76

 Bros

0.75

Activations Density 0.008%

the name "Leonard" with varying levels of activation

the name "Leonard" in various contexts

No Comments

No Known Activations