INDEX

Explanations

phrases related to being enclosed or surrounded by something

oai_token-act-pair · gpt-3.5-turbo

variations of the suffix "ed" or "ing" that indicate actions or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

itars

-0.70

issance

-0.64

ithmetic

-0.62

 partName

-0.62

entanyl

-0.61

onde

-0.61

=-=-=-=-

-0.60

imil

-0.59

ller

-0.58

ERG

-0.57

POSITIVE LOGITS

 confines

0.81

packed

0.79

puff

0.77

Layer

0.71

bub

0.71

river

0.70

ktop

0.68

 confinement

0.67

 insulated

0.66

Initialized

0.66

Activations Density 0.176%

phrases related to being enclosed or surrounded by something

variations of the suffix "ed" or "ing" that indicate actions or conditions

No Comments

No Known Activations