INDEX

Explanations

adjectives or terms representing a wide or extensive scope

oai_token-act-pair · gpt-3.5-turbo

references to the concept of "broad" 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

 Tycoon

-0.86

 Keeper

-0.79

ICAN

-0.76

LV

-0.74

 Disciple

-0.73

 Camer

-0.71

 Mistress

-0.68

agate

-0.67

Dialogue

-0.67

aminer

-0.65

POSITIVE LOGITS

sheet

1.46

sheets

1.30

leaf

1.11

ening

1.11

sword

1.01

 spectrum

0.98

 strokes

0.95

ened

0.90

side

0.88

 swath

0.86

Activations Density 0.022%

adjectives or terms representing a wide or extensive scope

references to the concept of "broad" in various contexts

No Comments

No Known Activations