INDEX

Explanations

JSON-related terms and formats

oai_token-act-pair · gpt-3.5-turbo

references to data serialization formats like JSON and XML

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rug

-0.73

ichita

-0.72

ggies

-0.71

Mum

-0.68

 Whitman

-0.66

eworld

-0.65

lain

-0.65

ners

-0.64

________________

-0.63

qua

-0.63

POSITIVE LOGITS

JSON

0.97

 JSON

0.96

RPC

0.92

 schema

0.87

©¶æ¥µ

0.87

 json

0.84

json

0.83

icter

0.82

 parser

0.79

Parser

0.72

Activations Density 0.008%

JSON-related terms and formats

references to data serialization formats like JSON and XML

No Comments

No Known Activations