INDEX

Explanations

mentions of job positions or ranks

oai_token-act-pair · gpt-3.5-turbo

references to different levels or grades within a ranking or classification 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

vous

-0.86

 Lucia

-0.80

Frames

-0.73

pload

-0.72

lez

-0.71

artney

-0.70

 Norris

-0.69

quad

-0.69

 Tycoon

-0.68

inos

-0.68

POSITIVE LOGITS

ership

0.86

 representation

0.78

 guidance

0.76

 equivalents

0.72

 advancement

0.71

 earners

0.70

 elevation

0.70

 treason

0.70

level

0.68

 managerial

0.68

Activations Density 0.018%

mentions of job positions or ranks

references to different levels or grades within a ranking or classification context

No Comments

No Known Activations