INDEX

Explanations

specific conditions or outcomes related to actions taken or not taken

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

%);

-0.63

%).

-0.57

çīĪ

-0.56

================================================================

-0.55

 Cosponsors

-0.54

.}

-0.53

chieve

-0.53

.''.

-0.53

aspberry

-0.53

©¶æ¥µ

-0.52

POSITIVE LOGITS

 then

1.02

 they

0.99

 anymore

0.96

 THEN

0.92

 please

0.91

it

0.88

you

0.88

then

0.87

 tomorrow

0.86

 sooner

0.84

Activations Density 0.722%

specific conditions or outcomes related to actions taken or not taken

No Comments

No Known Activations