INDEX

Explanations

statements expressing preference or intention

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

ĳå£«

-0.67

OTOS

-0.65

porter

-0.64

none

-0.63

court

-0.63

 Dres

-0.61

teness

-0.61

 Bland

-0.60

txt

-0.59

 Daniels

-0.59

POSITIVE LOGITS

 fared

1.09

 cope

0.95

 handled

0.94

 perce

0.88

 approached

0.87

 relates

0.86

 perceive

0.85

 reacted

0.85

 fare

0.85

 handles

0.85

Activations Density 0.106%

statements expressing preference or intention

No Comments

No Known Activations