INDEX

Explanations

assertions of truth, often followed by additional context or information

oai_token-act-pair · gpt-3.5-turbo

statements asserting truths or facts

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

ebus

-0.82

onding

-0.67

 arms

-0.67

reciation

-0.64

ido

-0.64

 erection

-0.63

 Quit

-0.62

robe

-0.61

pload

-0.60

 Maintenance

-0.59

POSITIVE LOGITS

etheless

0.80

©¶æ

0.68

Roe

0.66

 fiction

0.64

anuts

0.61

plain

0.61

 Fargo

0.60

cerning

0.59

matter

0.59

Bloom

0.59

Activations Density 0.170%

assertions of truth, often followed by additional context or information

statements asserting truths or facts

No Comments

No Known Activations