INDEX

Explanations

words related to marking or making a mark

oai_token-act-pair · gpt-3.5-turbo

references to "marks" or "marking" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

ILLE

-0.78

Kut

-0.75

rouch

-0.70

 traged

-0.67

annels

-0.66

urus

-0.66

anooga

-0.66

agan

-0.64

eger

-0.63

 Chrys

-0.62

POSITIVE LOGITS

manship

1.50

down

1.14

downs

1.10

eters

1.00

 Twain

0.94

ups

0.92

emark

0.91

posts

0.88

eting

0.88

marks

0.85

Activations Density 0.033%

words related to marking or making a mark

references to "marks" or "marking" in various contexts

No Comments

No Known Activations