INDEX

Explanations

phrases related to receiving or making phone calls

oai_token-act-pair · gpt-3.5-turbo

instances of the word "call" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 unden

-0.76

 istg

-0.74

 concentration

-0.72

 exha

-0.70

inth

-0.69

 stocking

-0.67

 bunny

-0.64

 Flavoring

-0.64

inh

-0.63

 vagina

-0.62

POSITIVE LOGITS

call

1.26

backs

1.23

igraph

1.00

 call

1.00

 Call

0.98

Call

0.96

0.93

 calls

0.92

 Niet

0.90

 CALL

0.87

Activations Density 0.037%

phrases related to receiving or making phone calls

instances of the word "call" in various contexts

No Comments

No Known Activations