INDEX

Explanations

explicit mentions of "specific" things or topics

oai_token-act-pair · gpt-3.5-turbo

phrases indicating specificity in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rican

-0.72

OWER

-0.71

mol

-0.67

 Feldman

-0.65

-0.64

 Springer

-0.64

 Dinosaur

-0.62

http

-0.61

 Neighbor

-0.61

 Hilton

-0.61

POSITIVE LOGITS

ities

1.07

ally

1.01

ivity

0.92

itarian

0.91

arily

0.90

iveness

0.88

ivities

0.88

ality

0.83

iations

0.83

atively

0.82

Activations Density 0.007%

explicit mentions of "specific" things or topics

phrases indicating specificity in context

No Comments

No Known Activations