INDEX

Explanations

terms related to being prepared or ready

oai_token-act-pair · gpt-3.5-turbo

contextual references to being prepared or set for action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 monetary

-0.66

 massage

-0.65

 tort

-0.64

 paras

-0.64

 unexplained

-0.62

 causes

-0.61

 torment

-0.61

 Investigator

-0.61

 Malt

-0.61

 Charm

-0.60

POSITIVE LOGITS

ready

4.62

Ready

2.16

 Ready

1.65

 ready

1.64

 readiness

1.16

done

1.11

friendly

1.06

eligible

1.04

worthy

1.03

readable

1.02

Activations Density 0.010%

terms related to being prepared or ready

contextual references to being prepared or set for action

No Comments

No Known Activations