INDEX

Explanations

descriptions of different states or conditions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "state" and its variations in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Flavoring

-0.94

poons

-0.83

anos

-0.79

Maw

-0.76

emis

-0.73

bats

-0.72

 Aires

-0.71

peria

-0.70

edin

-0.68

ogle

-0.67

POSITIVE LOGITS

 legislatures

0.96

 legislature

0.87

 state

0.83

wide

0.81

SPONSORED

0.81

less

0.80

state

0.79

rooms

0.78

lessness

0.76

room

0.76

Activations Density 0.050%

descriptions of different states or conditions

instances of the word "state" and its variations in different contexts

No Comments

No Known Activations