INDEX

Explanations

the word "off" used figuratively, typically in the context of severe consequences or actions

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

MAT

-0.79

FUL

-0.74

WM

-0.72

older

-0.69

 Frie

-0.67

YN

-0.67

enegger

-0.66

geries

-0.63

ollen

-0.63

antine

-0.62

POSITIVE LOGITS

 limbs

0.88

 contact

0.79

pring

0.78

 limb

0.78

ricular

0.78

 communication

0.76

 access

0.75

 valves

0.75

 abruptly

0.73

 disbelief

0.72

Activations Density 0.030%

the word "off" used figuratively, typically in the context of severe consequences or actions

No Comments

No Known Activations