INDEX

Explanations

phrases indicating capability or capacity to deal with something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "handle" 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 @ 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

urt

-0.76

ester

-0.76

esis

-0.75

ischer

-0.71

elong

-0.71

hemat

-0.69

laws

-0.69

owicz

-0.67

esters

-0.67

eny

-0.67

POSITIVE LOGITS

bars

1.14

 handles

0.86

 handle

0.81

ttes

0.73

 Handle

0.68

 grips

0.64

ª

0.64

 handling

0.64

IVER

0.63

atives

0.63

Activations Density 0.015%

phrases indicating capability or capacity to deal with something

occurrences of the word "handle" in various contexts

No Comments

No Known Activations