INDEX

Explanations

references to the name "Karl"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Karl" along with associated figures or terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ledged

-0.62

BLE

-0.61

âĢ¢âĢ¢

-0.60

 agency

-0.60

 precisely

-0.58

 Scotia

-0.58

cale

-0.57

 cleaner

-0.57

llan

-0.54

ointed

-0.53

POSITIVE LOGITS

 Rove

1.40

ovy

1.09

sson

1.05

Karl

1.05

 Marx

0.92

etics

0.90

stad

0.87

sen

0.86

stadt

0.85

rade

0.83

Activations Density 0.034%

references to the name "Karl"

mentions of the name "Karl" along with associated figures or terms

No Comments

No Known Activations