INDEX

Explanations

words related to the physical object "cup"

oai_token-act-pair · gpt-3.5-turbo

references to cups

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 targets

-0.68

Mes

-0.67

 targeting

-0.66

 targeted

-0.63

 executed

-0.62

Tel

-0.61

 Warn

-0.61

 looting

-0.60

 Shar

-0.60

 serial

-0.59

POSITIVE LOGITS

cup

4.66

cup

1.34

cream

1.33

 cups

1.25

bowl

1.24

cu

1.18

Cup

1.17

beer

1.15

 Cups

1.06

pants

1.05

Activations Density 0.019%

words related to the physical object "cup"

references to cups

No Comments

No Known Activations