INDEX

Explanations

phrases related to actions or processes that involve a sequence of steps

oai_token-act-pair · gpt-3.5-turbo

procedural instructions or steps in a process

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 contrary

-0.72

paren

-0.72

disabled

-0.68

 contradicted

-0.66

anity

-0.66

¥µ

-0.65

angering

-0.65

esity

-0.65

atro

-0.65

Russ

-0.64

POSITIVE LOGITS

 then

1.13

 optionally

1.07

 assigns

1.01

 determines

0.99

 prest

0.97

 whichever

0.95

 assign

0.92

 executes

0.92

 THEN

0.91

 sends

0.91

Activations Density 0.641%

phrases related to actions or processes that involve a sequence of steps

procedural instructions or steps in a process

No Comments

No Known Activations