INDEX

Explanations

locations and event information

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

Data Type

torch.float32

Hook Point

blocks.8.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

 grounding

-0.86

 induct

-0.74

 sneak

-0.71

 ambush

-0.71

 marked

-0.71

 chained

-0.70

 overlooked

-0.69

 sway

-0.68

 scaling

-0.68

 accent

-0.68

POSITIVE LOGITS

com

1.72

org

1.65

edu

1.45

net

1.45

blogspot

1.42

exe

1.41

wordpress

1.39

tumblr

1.38

gov

1.33

dll

1.30

Activations Density 0.926%

locations and event information

No Comments

No Known Activations