INDEX

Explanations

instances where something is clearly described or evident

oai_token-act-pair · gpt-3.5-turbo

explicit indications of certainty or clarity

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

lÃ©

-0.72

rey

-0.71

lav

-0.70

oleon

-0.69

aily

-0.69

ucky

-0.68

ourke

-0.68

RY

-0.68

imer

-0.67

uese

-0.66

POSITIVE LOGITS

 deline

0.96

 distinguish

0.91

 identifiable

0.86

 differentiated

0.84

 marked

0.83

 spelled

0.81

 differentiate

0.81

 distinguishing

0.79

 outwe

0.79

 distinguishes

0.78

Activations Density 0.017%

instances where something is clearly described or evident

explicit indications of certainty or clarity

No Comments

No Known Activations