INDEX

Explanations

references to the term "gan" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to individuals named "gan."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 differential

-0.74

İĭ

-0.67

 mates

-0.62

 authoritative

-0.61

PER

-0.60

NL

-0.59

 prejudice

-0.58

Spanish

-0.58

 pity

-0.57

 fright

-0.57

POSITIVE LOGITS

igans

1.13

igan

0.98

gan

0.97

culosis

0.97

gha

0.94

ciating

0.90

ãĥĥãĤ¯

0.89

cest

0.87

omal

0.87

ecake

0.86

Activations Density 0.007%

references to the term "gan" in various contexts

references to individuals named "gan."

No Comments

No Known Activations