INDEX

Explanations

concepts related to effort and emotional intensity

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

ħĭ

-0.74

ãĤ´ãĥ³

-0.71

xit

-0.69

-0.63

CHAT

-0.59

 Continued

-0.59

 Avenger

-0.57

 interpreter

-0.56

 operative

-0.56

riter

-0.55

POSITIVE LOGITS

 nowadays

0.87

 lately

0.76

(~

0.73

(>

0.72

 crammed

0.72

related

0.70

aic

0.68

 wasted

0.64

ðŁĳ

0.64

 upfront

0.64

Activations Density 0.139%

concepts related to effort and emotional intensity

No Comments

No Known Activations