INDEX

Explanations

legal terms and legislative 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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

auga

-0.75

ument

-0.69

aukee

-0.64

okia

-0.62

 Begins

-0.61

ospace

-0.61

roid

-0.60

\-

-0.59

ãĤº

-0.57

POSITIVE LOGITS

neath

1.09

lining

1.05

dogs

1.04

cuts

1.03

pins

1.03

lines

0.99

cutting

0.99

lying

0.98

lie

0.97

graduate

0.97

Activations Density 0.059%

legal terms and legislative information

No Comments

No Known Activations