INDEX

Explanations

uncommon or rare terms in statistical data or technical documents

oai_token-act-pair · gpt-3.5-turbo

terms related to rarity and frequency classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

choes

-0.93

Ń·

-0.83

pora

-0.80

ailable

-0.80

ebin

-0.79

cients

-0.74

galitarian

-0.72

nikov

-0.71

olulu

-0.69

utra

-0.69

POSITIVE LOGITS

ly

0.94

ensical

0.87

Parameters

0.80

wealth

0.80

Law

0.78

 Lisp

0.74

 Common

0.74

JS

0.73

abbage

0.73

Sense

0.71

Activations Density 0.011%

uncommon or rare terms in statistical data or technical documents

terms related to rarity and frequency classifications

No Comments

No Known Activations