INDEX

Explanations

phrases related to expressing strong opinions or directives

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

uthor

-0.83

OOL

-0.74

ESE

-0.70

DAQ

-0.70

natureconservancy

-0.69

ctic

-0.67

soever

-0.66

 Seym

-0.65

HER

-0.64

ÄŁ

-0.64

POSITIVE LOGITS

blank

0.94

iasis

0.89

point

0.86

 Reyes

0.85

 blank

0.84

lessly

0.83

points

0.81

 deduction

0.80

 pointing

0.78

entin

0.76

Activations Density 4.807%

phrases related to expressing strong opinions or directives

No Comments

No Known Activations