INDEX

Explanations

special characters and symbols followed by a colon

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

 conversions

-0.77

 constit

-0.75

 conversion

-0.69

 divest

-0.67

 corpus

-0.66

 distinguished

-0.65

 seiz

-0.65

 membership

-0.65

 succession

-0.65

 lifetime

-0.64

POSITIVE LOGITS

-)

1.24

lol

1.13

/)

1.12

_>

1.08

)"

1.08

laugh

1.08

DD

1.07

rolley

1.03

1.02

*=-

1.01

Activations Density 0.081%

special characters and symbols followed by a colon

No Comments

No Known Activations