INDEX

Explanations

instances where something is being represented or portrayed in a particular way

oai_token-act-pair · gpt-3.5-turbo

references to the concept of reflection or things that embody reflection

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Klu

-0.65

Hun

-0.65

killer

-0.64

oen

-0.63

Word

-0.62

 Mellon

-0.61

 Mechan

-0.61

opolis

-0.61

Pul

-0.60

illa

-0.59

POSITIVE LOGITS

 reflect

3.68

reflect

2.27

 reflects

2.18

 reflecting

2.17

 Reflect

2.15

 reflected

2.08

 reflective

2.06

 reflection

1.82

 reflections

1.45

 mirror

1.43

Activations Density 0.008%

instances where something is being represented or portrayed in a particular way

references to the concept of reflection or things that embody reflection

No Comments

No Known Activations