INDEX

Explanations

phrases related to unexpected outcomes or consequences

oai_token-act-pair · gpt-3.5-turbo

phrases that include the expression "ended up."

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

ongyang

-0.76

 assurance

-0.68

heed

-0.67

mens

-0.62

raint

-0.60

men

-0.59

 Rothschild

-0.59

ï¸ı

-0.59

mouth

-0.59

 indication

-0.58

POSITIVE LOGITS

skirts

0.76

ymes

0.73

 TAMADRA

0.71

ãĥĩãĤ£

0.71

 coli

0.69

steen

0.68

 redes

0.66

acters

0.65

adesh

0.65

icably

0.64

Activations Density 0.032%

phrases related to unexpected outcomes or consequences

phrases that include the expression "ended up."

No Comments

No Known Activations