INDEX

Explanations

discussions related to government, law, and strategy

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

ANC

-0.77

ADVERTISEMENT

-0.72

ince

-0.72

 Older

-0.72

UC

-0.69

uta

-0.69

MK

-0.67

ongo

-0.67

uty

-0.66

NULL

-0.66

POSITIVE LOGITS

 THEN

1.40

 then

1.38

 thereby

1.14

 secondly

1.03

 thus

1.02

vo

1.01

romeda

0.99

 hopefully

0.97

 prest

0.97

 reap

0.96

Activations Density 1.858%

discussions related to government, law, and strategy

No Comments

No Known Activations

discussions related to government, law, and strategy

No Comments

No Known Activations