INDEX

Explanations

determiners followed by a form of 'to be' and adjectives

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "to be" in various forms and 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

 DEFENSE

-0.72

 Pants

-0.66

criminal

-0.65

SHIP

-0.64

 Mellon

-0.63

arna

-0.63

edom

-0.60

 Dayton

-0.60

ince

-0.59

aii

-0.58

POSITIVE LOGITS

 wont

0.78

 attest

0.73

 hitherto

0.73

[|

0.72

 evidenced

0.72

 dictated

0.69

 regenerate

0.67

 progressed

0.65

 presently

0.64

 often

0.64

Activations Density 0.225%

determiners followed by a form of 'to be' and adjectives

instances of the verb "to be" in various forms and contexts

No Comments

No Known Activations