INDEX

Explanations

words related to cause and effect or reasons

oai_token-act-pair · gpt-3.5-turbo

transitional phrases and causal connections in argumentation

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ´ãĥ³

-1.07

Released

-0.84

Synopsis

-0.83

 exting

-0.75

Reviewed

-0.75

hig

-0.75

Born

-0.74

Winged

-0.70

ãĤ´

-0.69

tyard

-0.69

POSITIVE LOGITS

we

1.32

 somebody

1.21

 everybody

1.20

1.17

 anybody

1.13

 there

1.09

 they

1.01

 yeah

0.98

you

0.98

if

0.97

Activations Density 0.520%

words related to cause and effect or reasons

transitional phrases and causal connections in argumentation

No Comments

No Known Activations