INDEX

Explanations

phrases related to news reporting on events and incidents

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

achus

-0.86

terness

-0.75

ppe

-0.74

cation

-0.73

zzle

-0.73

onis

-0.72

let

-0.71

Ã«

-0.70

iness

-0.70

osate

-0.70

POSITIVE LOGITS

 kinds

1.13

 sorts

1.09

 fellows

0.98

 facts

0.97

 types

0.96

 truths

0.95

 guys

0.94

 shores

0.87

 gentlemen

0.86

 occurrences

0.86

Activations Density 0.272%

phrases related to news reporting on events and incidents

No Comments

No Known Activations