INDEX

Explanations

prepositions followed by a place or event

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

Introduced

-0.68

ieth

-0.66

 nicknamed

-0.66

 survives

-0.65

 persisted

-0.65

NT

-0.63

 approached

-0.63

 BELOW

-0.62

 deserves

-0.61

Ohio

-0.61

POSITIVE LOGITS

gotten

0.84

aging

0.83

rocal

0.83

WARD

0.82

ked

0.82

 holidays

0.80

 example

0.80

 awhile

0.80

 lunch

0.79

geries

0.79

Activations Density 0.127%

prepositions followed by a place or event

No Comments

No Known Activations