INDEX

Explanations

phrases regarding treatment or conditions causing suffering

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

MQ

-0.72

 Sheet

-0.69

 Thoughts

-0.66

Ruk

-0.61

ono

-0.60

onday

-0.59

rake

-0.59

idays

-0.59

raid

-0.58

 sitting

-0.58

POSITIVE LOGITS

 resembles

1.04

 horr

1.04

 enables

0.89

 satisfies

0.81

 allows

0.81

 enhances

0.81

 exceeds

0.80

 corresponds

0.79

 resemble

0.79

 mirrors

0.78

Activations Density 0.173%

phrases regarding treatment or conditions causing suffering

No Comments

No Known Activations