INDEX

Explanations

statements involving someone telling information about themselves or others

oai_token-act-pair · gpt-3.5-turbo

statements of past actions or experiences involving individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Nationwide

-0.64

 Mandatory

-0.63

Translation

-0.62

stakes

-0.61

ormal

-0.59

 Header

-0.59

 Regions

-0.59

nesses

-0.58

 Globe

-0.58

duc

-0.57

POSITIVE LOGITS

 regretted

1.27

 regrets

1.21

'd

1.05

 slept

1.04

 hadn

1.02

 dreamed

0.98

 woke

0.96

 wished

0.95

 hates

0.94

 remembers

0.94

Activations Density 0.185%

statements involving someone telling information about themselves or others

statements of past actions or experiences involving individuals

No Comments

No Known Activations