INDEX

Explanations

references to different classes or classifications

oai_token-act-pair · gpt-3.5-turbo

references to classification levels or categories related to offenses or entities

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

issance

-0.79

OPLE

-0.70

 destro

-0.70

 guiActiveUn

-0.68

 ende

-0.68

 push

-0.68

orescence

-0.68

 skept

-0.65

 promise

-0.64

spr

-0.64

POSITIVE LOGITS

 Class

0.99

mates

0.99

Loader

0.95

mate

0.94

ifiers

0.93

Class

0.89

Cast

0.89

 Classes

0.88

ifier

0.87

Construct

0.84

Activations Density 0.011%

references to different classes or classifications

references to classification levels or categories related to offenses or entities

No Comments

No Known Activations