INDEX

Explanations

words related to adjustment or alignment

oai_token-act-pair · gpt-3.5-turbo

instances of the word "align" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ilk

-0.92

HAEL

-0.77

 Roof

-0.75

chal

-0.74

Lind

-0.70

sein

-0.70

Ox

-0.69

ek

-0.67

Kay

-0.66

je

-0.65

POSITIVE LOGITS

 aligned

1.20

ments

1.14

iances

1.12

 align

1.11

 alignment

1.05

arity

1.00

aligned

0.99

 alliances

0.91

align

0.90

 guiActiveUn

0.87

Activations Density 0.009%

words related to adjustment or alignment

instances of the word "align" and its variations

No Comments

No Known Activations