INDEX

Explanations

mentions of specific locations or entities that are currently involved in various events

oai_token-act-pair · gpt-3.5-turbo

punctuation and statements indicating ongoing discussions or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 outwe

-0.74

 afterlife

-0.73

 regener

-0.71

 unus

-0.66

 enjoyment

-0.65

 immortality

-0.65

livion

-0.64

 compost

-0.64

 intangible

-0.64

gment

-0.64

POSITIVE LOGITS

 Yesterday

1.26

Yesterday

1.22

Among

1.17

These

1.15

 These

1.03

Reports

1.02

Recently

1.02

Documents

1.02

 Among

1.01

 Recently

1.00

Activations Density 0.539%

mentions of specific locations or entities that are currently involved in various events

punctuation and statements indicating ongoing discussions or actions

No Comments

No Known Activations