INDEX

Explanations

speech-related terms or phrases

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

 Soldiers

-0.78

 Associ

-0.65

 Restaur

-0.63

 Leone

-0.61

 Ryder

-0.61

hews

-0.60

 Dani

-0.60

xon

-0.60

 Mirage

-0.59

Ã±o

-0.59

POSITIVE LOGITS

 imped

1.03

writer

0.90

speech

0.90

writing

0.88

 uttered

0.87

writers

0.86

 speech

0.82

anooga

0.78

 transcripts

0.78

press

0.77

Activations Density 0.041%

speech-related terms or phrases

No Comments

No Known Activations