INDEX

Explanations

terms related to remaining in a specific state or location

oai_token-act-pair · gpt-3.5-turbo

instances of the word "stay" in various contexts

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

NetMessage

-0.83

Ñı

-0.70

DIT

-0.68

 Cosponsors

-0.68

 Parables

-0.67

sha

-0.66

ologies

-0.66

tained

-0.65

oros

-0.63

itivity

-0.62

POSITIVE LOGITS

 afloat

0.91

overs

0.81

cation

0.79

 silent

0.76

 vigil

0.72

 awake

0.71

 alive

0.71

 quiet

0.70

 undefeated

0.69

emate

0.69

Activations Density 0.020%

terms related to remaining in a specific state or location

instances of the word "stay" in various contexts

No Comments

No Known Activations