INDEX

Explanations

adjectives and phrases related to evaluation or comparison

oai_token-act-pair · gpt-3.5-turbo

concepts related to distinctions between theory and fact or reality

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Ĭ±

-0.73

 Cosponsors

-0.69

 Divide

-0.65

bryce

-0.64

 Decre

-0.64

apego

-0.63

thora

-0.62

©¶æ¥µ

-0.61

 Delay

-0.61

å¦

-0.61

POSITIVE LOGITS

 actual

1.99

 real

1.51

 reality

1.43

 Actual

1.35

actual

1.32

 tangible

1.24

 actually

1.21

 realities

1.21

 genuine

1.20

real

1.19

Activations Density 0.649%

adjectives and phrases related to evaluation or comparison

concepts related to distinctions between theory and fact or reality

No Comments

No Known Activations