INDEX

Explanations

words related to product features or characteristics

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

 tremend

-0.68

 Chimera

-0.68

 consolidation

-0.66

ingham

-0.63

 alloc

-0.62

Requirements

-0.60

 Anthem

-0.59

milo

-0.58

 manufact

-0.58

 Legions

-0.58

POSITIVE LOGITS

ghan

1.39

lder

1.11

ze

1.07

gee

0.97

fter

0.95

xt

0.94

hee

0.94

qs

0.91

vern

0.91

vel

0.90

Activations Density 0.070%

words related to product features or characteristics

No Comments

No Known Activations