INDEX

Explanations

phrases where something is being explained or discussed in detail

oai_token-act-pair · gpt-3.5-turbo

occurrences of the words "described" and "discussed."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

idity

-0.80

cot

-0.73

lev

-0.70

ajo

-0.70

riot

-0.67

opers

-0.66

adra

-0.65

otion

-0.65

ionage

-0.65

bal

-0.65

POSITIVE LOGITS

ĸļ

0.86

escription

0.83

 Parenthood

0.82

 supra

0.76

 inconsist

0.76

ãĤ¶

0.73

 above

0.72

ktop

0.71

 herein

0.68

ocument

0.68

Activations Density 0.180%

phrases where something is being explained or discussed in detail

occurrences of the words "described" and "discussed."

No Comments

No Known Activations