INDEX

Explanations

statements or judgments of correctness

oai_token-act-pair · gpt-3.5-turbo

statements about accuracy or correctness

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

aden

-0.74

GGGGGGGG

-0.74

 Valhalla

-0.72

EMOTE

-0.72

CHO

-0.66

neys

-0.63

ILY

-0.63

thin

-0.62

doms

-0.62

Connector

-0.61

POSITIVE LOGITS

ives

0.95

eous

0.86

fully

0.85

ibly

0.84

 guiActiveUn

0.80

 answers

0.78

ible

0.77

aber

0.75

 translations

0.73

correct

0.73

Activations Density 0.015%

statements or judgments of correctness

statements about accuracy or correctness

No Comments

No Known Activations