INDEX

Explanations

phrases related to diminishing or decreasing qualities or conditions

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

è¦ļéĨĴ

-0.73

ãĤ±

-0.68

gow

-0.66

arov

-0.64

ãĥīãĥ©

-0.64

square

-0.63

eering

-0.61

 Investigative

-0.58

RECT

-0.58

OH

-0.58

POSITIVE LOGITS

 abruptly

0.90

 waning

0.87

 peacefully

0.81

 dwindling

0.79

 extinguished

0.78

 fading

0.74

ldom

0.74

orph

0.74

 fade

0.73

 fades

0.73

Activations Density 0.101%

phrases related to diminishing or decreasing qualities or conditions

No Comments

No Known Activations