INDEX

Explanations

parentheses and numerical information

oai_token-act-pair · gpt-3.5-turbo

open and closed parentheses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 nuts

-0.71

nesday

-0.68

wagen

-0.68

 lull

-0.66

pheus

-0.62

 orche

-0.62

 Unic

-0.60

 inund

-0.59

 absor

-0.59

 nearest

-0.59

POSITIVE LOGITS

Reuters

0.98

emphasis

0.97

credit

0.97

...)

0.95

?)

0.94

aka

0.93

click

0.92

formerly

0.90

please

0.90

Credit

0.89

Activations Density 0.109%

parentheses and numerical information

open and closed parentheses

No Comments

No Known Activations