INDEX

Explanations

verbs related to desire or wanting

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

-0.73

icol

-0.66

livious

-0.62

trust

-0.61

semble

-0.60

unny

-0.60

icist

-0.60

rir

-0.59

ulty

-0.58

errors

-0.57

POSITIVE LOGITS

 revenge

0.91

reprene

0.90

to

0.87

 answers

0.77

 nothing

0.74

 attention

0.74

 clarification

0.73

 assurances

0.72

 someone

0.72

 something

0.72

Activations Density 0.943%

verbs related to desire or wanting

No Comments

No Known Activations