INDEX

Explanations

passwords or codes indicated by the characters "WD" followed by a number

oai_token-act-pair · gpt-3.5-turbo

references to a specific identifier or term related to an entity or category

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

odies

-0.86

auld

-0.83

ussian

-0.76

ylon

-0.75

berus

-0.73

auldron

-0.69

emetery

-0.69

zers

-0.68

 Adviser

-0.68

arde

-0.68

POSITIVE LOGITS

ership

0.89

ictionary

0.82

irection

0.78

ansas

0.77

ski

0.76

ition

0.75

RECT

0.74

0.71

emonic

0.71

ication

0.70

Activations Density 0.024%

passwords or codes indicated by the characters "WD" followed by a number

references to a specific identifier or term related to an entity or category

No Comments

No Known Activations