INDEX

Explanations

mentions of the word "Med", often related to proper nouns like "Medellin" or "medical"

oai_token-act-pair · gpt-3.5-turbo

references to the term "Med" in various contexts

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

chnology

-0.84

 innocence

-0.75

ģĸ

-0.68

holding

-0.67

ãģį

-0.67

 Wilde

-0.65

ngth

-0.65

!/

-0.63

AUT

-0.63

roud

-0.63

POSITIVE LOGITS

ications

1.15

iterranean

1.13

allion

1.05

itations

1.05

ication

1.02

iator

0.99

iation

0.96

itated

0.95

ient

0.94

usa

0.93

Activations Density 0.017%

mentions of the word "Med", often related to proper nouns like "Medellin" or "medical"

references to the term "Med" in various contexts

No Comments

No Known Activations