INDEX

Explanations

specific phrases starting with the word "From" and followed by particular topics or contexts

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

 respons

-0.79

 manif

-0.73

 eleph

-0.70

 colle

-0.70

nesday

-0.68

ilaterally

-0.66

 toget

-0.66

akespe

-0.66

cc

-0.64

onest

-0.64

POSITIVE LOGITS

 Expand

0.85

 largeDownload

0.82

 Detected

0.80

Associated

0.76

 SHARES

0.75

Streamer

0.75

 ILCS

0.73

 Finder

0.73

Nintendo

0.72

CLUS

0.72

Activations Density 3.057%

specific phrases starting with the word "From" and followed by particular topics or contexts

No Comments

No Known Activations