INDEX

Explanations

words related to movement or action

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

qua

-0.76

hetical

-0.74

vi

-0.72

ged

-0.69

jo

-0.68

igh

-0.66

ISA

-0.66

encies

-0.66

iPhone

-0.65

BT

-0.65

POSITIVE LOGITS

 corners

1.05

 town

0.96

town

0.76

 freely

0.71

 corner

0.69

 uncontroll

0.68

 frantically

0.67

 Amsterdam

0.67

 perimeter

0.66

 maniac

0.65

Activations Density 0.036%

words related to movement or action

No Comments

No Known Activations