INDEX

Explanations

phrases expressing positive or negative evaluations and judgments

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

\'

-0.76

onut

-0.66

ench

-0.65

ohyd

-0.64

kee

-0.62

acca

-0.62

ilian

-0.61

ivalry

-0.59

haw

-0.58

cies

-0.58

POSITIVE LOGITS

 someday

0.83

 tomorrow

0.81

if

0.78

Osw

0.69

 Wouldn

0.68

 sooner

0.67

 wiser

0.67

morrow

0.66

 feas

0.65

 forever

0.65

Activations Density 0.191%

phrases expressing positive or negative evaluations and judgments

No Comments

No Known Activations