INDEX

Explanations

phrases related to security and privacy concerns

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ilet

-0.67

gently

-0.63

 confidently

-0.62

 Calories

-0.61

UGH

-0.60

GES

-0.59

resents

-0.58

rahim

-0.57

Together

-0.56

ITH

-0.56

POSITIVE LOGITS

 cumbers

0.78

 constraints

0.77

 circumstances

0.76

_.

0.73

.ï¿½

0.73

 surrounding

0.72

 inexper

0.70

 nature

0.70

iHUD

0.70

 conflic

0.70

Activations Density 0.437%

phrases related to security and privacy concerns

No Comments

No Known Activations