INDEX

Explanations

names of specific individuals, possibly related to various activities or events

oai_token-act-pair · gpt-3.5-turbo

references to the word "den."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

uana

-0.79

accompan

-0.74

EED

-0.71

onies

-0.69

avorite

-0.68

 behav

-0.67

cgi

-0.67

olicy

-0.64

ROM

-0.64

DIV

-0.64

POSITIVE LOGITS

ning

1.00

wald

0.97

unci

0.95

izens

0.90

ãĤª

0.90

heimer

0.90

umerable

0.89

heim

0.89

ving

0.86

unciation

0.82

Activations Density 0.011%

names of specific individuals, possibly related to various activities or events

references to the word "den."

No Comments

No Known Activations