INDEX

Explanations

grammatical structures related to directives, orders, or commands

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "directive" 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

ock

-0.78

-0.76

oor

-0.73

 Alto

-0.72

 Blacks

-0.70

OWS

-0.69

arsity

-0.69

pocket

-0.68

Gordon

-0.67

ikes

-0.67

POSITIVE LOGITS

 directives

1.16

 directive

1.03

 decree

0.98

 Directive

0.85

 directs

0.81

nod

0.77

 workflow

0.76

 directing

0.76

 guiActiveUn

0.74

 choke

0.73

Activations Density 0.017%

grammatical structures related to directives, orders, or commands

occurrences of the word "directive" in various contexts

No Comments

No Known Activations