INDEX

Explanations

names of individuals being appointed or invited to specific roles or events

oai_token-act-pair · gpt-3.5-turbo

occurrences of the verb "be" in various contexts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥ¡

-0.74

 Saying

-0.73

 Being

-0.68

being

-0.64

 prompting

-0.63

 Matter

-0.62

 noticing

-0.62

 prediction

-0.62

 Struggle

-0.61

 confinement

-0.60

POSITIVE LOGITS

league

0.98

amed

0.96

headed

0.91

arers

0.91

 replaced

0.87

 interviewed

0.86

heading

0.85

head

0.85

 reunited

0.84

que

0.82

Activations Density 0.113%

names of individuals being appointed or invited to specific roles or events

occurrences of the verb "be" in various contexts

No Comments

No Known Activations