INDEX

Explanations

phrases that indicate actions, especially commands and suggestions

oai_token-act-pair · gpt-4o-mini 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

ELD

-0.79

eur

-0.65

-0.63

lessly

-0.60

 Associated

-0.57

 asset

-0.57

 haun

-0.57

Required

-0.57

è¦ļéĨĴ

-0.56

POSITIVE LOGITS

 assume

1.15

 suppose

1.08

 discuss

1.06

 pretend

1.05

 speculate

0.99

 summarize

0.96

 presume

0.95

 examine

0.93

 conclude

0.93

 celebrate

0.93

Activations Density 0.037%

phrases that indicate actions, especially commands and suggestions

No Comments

No Known Activations