INDEX

Explanations

keywords related to security, durability, and providing information

oai_token-act-pair · gpt-3.5-turbo

words related to security and safety mechanisms

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

 Journals

-0.64

 Revised

-0.64

 Burk

-0.61

 Reverend

-0.61

 disbanded

-0.61

wagon

-0.60

 Ancients

-0.58

ERC

-0.58

 Vend

-0.58

ç«

-0.57

POSITIVE LOGITS

 capability

0.87

 edge

0.83

iability

0.81

 superiority

0.81

 advantage

0.78

imum

0.77

 flexibility

0.75

 unparalleled

0.74

ability

0.74

 unmatched

0.74

Activations Density 0.586%

keywords related to security, durability, and providing information

words related to security and safety mechanisms

No Comments

No Known Activations