INDEX

Explanations

The neuron looks for mentions of a specific person or entity named "Bleeding Cool"

oai_token-act-pair · gpt-3.5-turbo

references to a specific brand or entity related to "Bleeding Cool"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

utive

-0.71

 motel

-0.65

 expense

-0.63

 orig

-0.63

CI

-0.63

 nexus

-0.63

USS

-0.61

gt

-0.61

 hatch

-0.61

cott

-0.61

POSITIVE LOGITS

Ble

4.08

Ble

3.15

 bleed

1.62

ble

1.45

 Bloody

1.35

bleacher

1.17

 bleeding

1.17

ble

1.15

 Bers

1.14

 Bleach

1.06

Activations Density 0.020%

The neuron looks for mentions of a specific person or entity named "Bleeding Cool"

references to a specific brand or entity related to "Bleeding Cool"

No Comments

No Known Activations