INDEX

Explanations

phrases related to certainty or possibility

oai_token-act-pair · gpt-3.5-turbo

the presence of the verb "be" in various forms

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

rones

-0.73

shall

-0.67

inav

-0.63

enaries

-0.61

culosis

-0.61

iasco

-0.60

 srfAttach

-0.59

 Telescope

-0.59

 Prest

-0.59

 guarantee

-0.58

POSITIVE LOGITS

 forgiven

1.00

 mistaken

0.96

 able

0.96

 construed

0.94

acons

0.85

 tempted

0.85

 regarded

0.83

 considered

0.83

league

0.82

 swayed

0.81

Activations Density 0.141%

phrases related to certainty or possibility

the presence of the verb "be" in various forms

No Comments

No Known Activations