INDEX

Explanations

programming-related words and terms related to syntax

oai_token-act-pair · gpt-3.5-turbo

references to programming language syntax

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

 Kore

-0.90

 Gund

-0.78

ECD

-0.78

birth

-0.77

ometown

-0.73

 Gamble

-0.72

 Gunn

-0.72

 Hanson

-0.71

IVERS

-0.68

 Bauer

-0.67

POSITIVE LOGITS

 syntax

1.21

ntax

1.14

hess

0.87

 notation

0.85

 tricks

0.83

 parser

0.81

 parsing

0.79

 language

0.79

 spelling

0.79

daq

0.78

Activations Density 0.015%

programming-related words and terms related to syntax

references to programming language syntax

No Comments

No Known Activations