INDEX

Explanations

the acronym "BRIAN" with varying levels of activation, possibly associated with specific contexts or entities

oai_token-act-pair · gpt-3.5-turbo

mentions of a specific organization or entity, likely related to research or initiatives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rules

-0.73

wagon

-0.71

holders

-0.67

forth

-0.67

hawk

-0.66

grass

-0.66

terson

-0.65

 sands

-0.63

kinson

-0.63

dayName

-0.63

POSITIVE LOGITS

ASON

1.03

KER

1.01

VEN

0.97

YA

0.92

HCR

0.92

EF

0.92

BE

0.91

etary

0.90

BLE

0.88

PT

0.88

Activations Density 0.009%

the acronym "BRIAN" with varying levels of activation, possibly associated with specific contexts or entities

mentions of a specific organization or entity, likely related to research or initiatives

No Comments

No Known Activations