INDEX

Explanations

details or descriptions in reports and agreements

oai_token-act-pair · gpt-3.5-turbo

verbs associated with detailed reporting and outlining of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

PI

-0.84

itus

-0.82

quer

-0.82

ogie

-0.78

iffe

-0.77

hene

-0.76

 Became

-0.73

veland

-0.72

soDeliveryDate

-0.71

soType

-0.70

POSITIVE LOGITS

 extensive

0.83

 numerous

0.76

 irregularities

0.76

 warnings

0.73

 inconsistencies

0.73

 inaccur

0.71

 deficiencies

0.71

 widespread

0.70

 unnecess

0.68

 sacrifices

0.68

Activations Density 0.346%

details or descriptions in reports and agreements

verbs associated with detailed reporting and outlining of information

No Comments

No Known Activations