INDEX

Explanations

mentions related to winning or achieving a high status or title

oai_token-act-pair · gpt-3.5-turbo

terms related to victory or champions 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 @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 reservations

-0.82

onica

-0.81

 Siber

-0.80

alle

-0.73

ogl

-0.72

xy

-0.72

 Debor

-0.71

asia

-0.71

zos

-0.71

eller

-0.70

POSITIVE LOGITS

interstitial

0.92

 reigning

0.91

puted

0.75

 crowned

0.74

 stagnant

0.71

Winner

0.69

 emperor

0.68

 pend

0.67

 vict

0.67

division

0.66

Activations Density 0.021%

mentions related to winning or achieving a high status or title

terms related to victory or champions in various contexts

No Comments

No Known Activations