INDEX

Explanations

actions related to unlocking or activating features or bonuses in a system or game

oai_token-act-pair · gpt-3.5-turbo

terms related to activation and unlocking in gaming or interactive experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.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

cas

-0.79

fare

-0.78

wig

-0.69

lli

-0.68

kr

-0.67

mos

-0.66

dding

-0.66

oss

-0.66

fur

-0.65

esan

-0.63

POSITIVE LOGITS

 dormant

0.83

 latent

0.76

 Unlock

0.75

 gates

0.74

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.74

 doors

0.74

CLASSIFIED

0.73

 abilities

0.71

uries

0.69

 recess

0.68

Activations Density 0.080%

actions related to unlocking or activating features or bonuses in a system or game

terms related to activation and unlocking in gaming or interactive experiences

No Comments

No Known Activations