INDEX

Explanations

phrases that present contrasting information or examples

oai_token-act-pair · gpt-3.5-turbo

conjunctions and transitional phrases suggesting relationships between ideas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

boro

-0.62

okane

-0.58

 doorway

-0.56

 clipboard

-0.55

 lockdown

-0.54

wordpress

-0.53

ideshow

-0.52

 spectator

-0.52

facade

-0.50

 picnic

-0.50

POSITIVE LOGITS

 however

1.14

 furthermore

1.12

 moreover

1.07

 therefore

1.01

Therefore

1.00

Moreover

0.99

Specifically

0.96

Furthermore

0.92

However

0.83

 Specifically

0.81

Activations Density 0.862%

phrases that present contrasting information or examples

conjunctions and transitional phrases suggesting relationships between ideas

No Comments

No Known Activations