INDEX

Explanations

mentions of the word "fine" used positively or in a complimentary context

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

omorphic

-0.85

omorph

-0.81

ulhu

-0.81

leaders

-0.76

raltar

-0.75

ilee

-0.75

 thwarted

-0.75

marine

-0.73

Hack

-0.71

riber

-0.70

POSITIVE LOGITS

 Gael

1.05

 tuning

1.02

 tuned

0.85

 tune

0.79

 fine

0.77

 linen

0.70

Fine

0.70

AY

0.69

MENTS

0.69

 Tune

0.67

Activations Density 0.854%

mentions of the word "fine" used positively or in a complimentary context

No Comments

No Known Activations