INDEX

Explanations

mentions of specific names, likely related to the name "Beng"

oai_token-act-pair · gpt-3.5-turbo

names of individuals, specifically those related to a sporting context

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

culosis

-0.74

ilial

-0.69

 Cause

-0.66

 fracture

-0.65

 fracturing

-0.65

eton

-0.65

 docking

-0.63

 sclerosis

-0.63

igham

-0.63

 dispens

-0.62

POSITIVE LOGITS

ober

0.81

 Frie

0.79

ħ

0.79

ī

0.76

¶

0.76

inx

0.75

ts

0.74

Ī

0.74

oul

0.73

AIN

0.73

Activations Density 0.039%

mentions of specific names, likely related to the name "Beng"

names of individuals, specifically those related to a sporting context

No Comments

No Known Activations