INDEX

Explanations

past tense verbs ending in 'ed' with a particular focus on negative events or actions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ivity

-0.68

å£

-0.66

LV

-0.65

 tabl

-0.65

 Parables

-0.63

 Forums

-0.63

 DRAG

-0.58

 Restaur

-0.56

ORY

-0.56

 Leilan

-0.55

POSITIVE LOGITS

up

0.73

coat

0.72

 down

0.71

ucket

0.70

down

0.70

lease

0.69

up

0.69

req

0.68

orah

0.67

athlon

0.67

Activations Density 0.086%

past tense verbs ending in 'ed' with a particular focus on negative events or actions

No Comments

No Known Activations