INDEX

Explanations

phrases or words related to the concept of correctness or accuracy

oai_token-act-pair · gpt-3.5-turbo

assertions of correctness or validity

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

EMOTE

-0.82

aden

-0.74

CHO

-0.73

doms

-0.70

 Valhalla

-0.70

lust

-0.69

SAY

-0.69

atos

-0.67

neys

-0.65

GGGGGGGG

-0.64

POSITIVE LOGITS

ives

0.90

correct

0.90

Correct

0.82

 answers

0.81

 spelling

0.81

orate

0.80

 guiActiveUn

0.80

 corrected

0.79

 answer

0.78

ively

0.74

Activations Density 0.006%

phrases or words related to the concept of correctness or accuracy

assertions of correctness or validity

No Comments

No Known Activations