INDEX

Explanations

the words "through" and indicates focus on the action of going through something

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

CTV

-0.79

abet

-0.76

isted

-0.76

Ring

-0.72

\/

-0.71

UTH

-0.68

etus

-0.68

IST

-0.65

juven

-0.62

oka

-0.62

POSITIVE LOGITS

 waivers

0.84

 intermedi

0.84

 puberty

0.76

 obscurity

0.76

inventoryQuantity

0.75

 adolescence

0.73

 hoops

0.72

 attrition

0.71

 adversity

0.71

 clenched

0.71

Activations Density 0.530%

the words "through" and indicates focus on the action of going through something

No Comments

No Known Activations