INDEX

Explanations

legal and judicial terms

oai_token-act-pair · gpt-3.5-turbo

terms related to authority or superiority

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

 Labrador

-0.64

gerald

-0.61

Ack

-0.58

 behold

-0.58

 drawing

-0.57

Syd

-0.57

REF

-0.57

 McCabe

-0.57

 Flan

-0.56

Cue

-0.56

POSITIVE LOGITS

antle

1.12

nants

1.10

ploy

1.06

nant

1.06

phasis

1.01

ovable

0.98

arkable

0.98

oval

0.95

phas

0.93

ixed

0.92

Activations Density 0.018%

legal and judicial terms

terms related to authority or superiority

No Comments

No Known Activations