INDEX

Explanations

mentions of a specific institution or location, specifically "Carleton."

oai_token-act-pair · gpt-3.5-turbo

references to specific educational institutions

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

gets

-0.77

acca

-0.77

elf

-0.76

fixes

-0.74

ging

-0.73

mond

-0.69

haw

-0.69

agger

-0.69

arcity

-0.69

boat

-0.68

POSITIVE LOGITS

gomery

0.82

 Lauder

0.80

leton

0.75

 Grove

0.71

 Pradesh

0.68

 blades

0.66

 Confederation

0.65

izabeth

0.65

Leaks

0.65

ANN

0.65

Activations Density 0.045%

mentions of a specific institution or location, specifically "Carleton."

references to specific educational institutions

No Comments

No Known Activations