INDEX

Explanations

words related to contrasting actions or decisions

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

andestine

-0.68

ENTS

-0.60

vez

-0.58

emate

-0.58

ankind

-0.57

 Shake

-0.57

 Smash

-0.57

gin

-0.56

ental

-0.56

iky

-0.56

POSITIVE LOGITS

 opting

0.91

 thereof

0.78

 preferring

0.77

opt

0.70

pling

0.70

ngth

0.69

 chose

0.69

ples

0.69

 choosing

0.67

 relying

0.67

Activations Density 0.240%

words related to contrasting actions or decisions

No Comments

No Known Activations