INDEX

Explanations

names or references to a specific object - "Jar"

oai_token-act-pair · gpt-3.5-turbo

references to jars

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ãģį

-0.74

ĨĴ

-0.74

ancial

-0.70

 Samoa

-0.69

éĹĺ

-0.68

EPA

-0.66

ierrez

-0.66

anooga

-0.66

eem

-0.65

 dividend

-0.65

POSITIVE LOGITS

Jar

1.23

Jar

1.14

assic

1.01

jar

0.95

 jars

0.94

rod

0.91

Ã³n

0.88

acca

0.86

thal

0.85

olla

0.83

Activations Density 0.009%

names or references to a specific object - "Jar"

references to jars

No Comments

No Known Activations