INDEX

Explanations

instances of the word "through" connecting entities or processes

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

\/

-0.77

etus

-0.73

abet

-0.73

CTV

-0.71

IST

-0.69

ãĥī

-0.68

ailable

-0.67

juven

-0.67

OST

-0.62

UTH

-0.62

POSITIVE LOGITS

 intermedi

1.00

 attrition

0.84

 waivers

0.83

 sheer

0.76

 channels

0.76

 clenched

0.75

 obscurity

0.72

 grape

0.71

 veins

0.71

put

0.68

Activations Density 1.097%

instances of the word "through" connecting entities or processes

No Comments

No Known Activations