INDEX

Explanations

phrases related to negative emotions or reactions towards statements 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

tein

-0.77

Ranked

-0.68

eatures

-0.67

entials

-0.64

ILA

-0.64

gravity

-0.62

allows

-0.58

ramid

-0.57

amins

-0.57

 Bermuda

-0.57

POSITIVE LOGITS

 aback

1.03

by

0.86

dy

0.78

 hearing

0.76

 enough

0.74

NESS

0.69

 seeing

0.68

 because

0.67

ragon

0.66

 wondering

0.66

Activations Density 0.099%

phrases related to negative emotions or reactions towards statements or actions

No Comments

No Known Activations