INDEX

Explanations

instructions or prompts to write

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ĭ±

-0.89

agara

-0.84

 Unsure

-0.67

 Ukrain

-0.66

allows

-0.66

rises

-0.66

nels

-0.65

negie

-0.64

arov

-0.64

EGA

-0.63

POSITIVE LOGITS

writing

0.89

itatively

0.86

smanship

0.85

 letters

0.84

 journal

0.84

lishing

0.83

 penned

0.80

 memos

0.80

writer

0.78

 essays

0.77

Activations Density 1.858%

instructions or prompts to write

No Comments

No Known Activations