INDEX

Explanations

words related to desires and intentions

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

VERTISEMENT

-0.64

icol

-0.64

anches

-0.61

livious

-0.60

icist

-0.59

semble

-0.58

trust

-0.58

rir

-0.58

ilian

-0.58

bug

-0.57

POSITIVE LOGITS

reprene

0.91

to

0.89

 revenge

0.82

 answers

0.76

 nothing

0.72

 assurances

0.72

 someone

0.69

 somebody

0.69

only

0.68

 attention

0.68

Activations Density 0.444%

words related to desires and intentions

No Comments

No Known Activations