INDEX

Explanations

the word "attempted"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "to" and related attempts or efforts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

minus

-0.69

eyes

-0.67

 Clouds

-0.65

 Needs

-0.64

liest

-0.64

ILY

-0.64

loads

-0.63

houses

-0.63

faces

-0.63

most

-0.62

POSITIVE LOGITS

 revive

1.21

 interfere

1.19

 replicate

1.18

 emulate

1.17

 recreate

1.16

 penetrate

1.15

 reconcile

1.14

 evade

1.12

 imitate

1.12

 defend

1.08

Activations Density 0.140%

the word "attempted"

occurrences of the word "to" and related attempts or efforts

No Comments

No Known Activations