INDEX

Explanations

information that is intentionally kept secret or not revealed

oai_token-act-pair · gpt-3.5-turbo

terms related to undisclosed or unspecified information and amounts

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

hedral

-0.76

alog

-0.75

builders

-0.74

Reviewer

-0.73

ranch

-0.69

=-=-=-=-=-=-=-=-

-0.67

 builders

-0.67

istries

-0.67

ickr

-0.66

workers

-0.66

POSITIVE LOGITS

 undisclosed

0.90

 amount

0.89

 unspecified

0.86

 amounts

0.84

 timeframe

0.70

TBD

0.69

TBA

0.67

 remainder

0.67

 quantity

0.66

 forthcoming

0.65

Activations Density 0.018%

information that is intentionally kept secret or not revealed

terms related to undisclosed or unspecified information and amounts

No Comments

No Known Activations