INDEX

Explanations

words related to actions or accomplishments

oai_token-act-pair · gpt-3.5-turbo

past tense verbs associated with completed 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arta

-0.72

yah

-0.69

entric

-0.68

wall

-0.64

copy

-0.64

veland

-0.64

ahon

-0.63

opus

-0.62

arel

-0.61

enta

-0.59

POSITIVE LOGITS

 yesterday

1.01

 last

0.90

 earlier

0.84

nesday

0.77

 originally

0.77

 unsuccessfully

0.77

 tremend

0.76

wolves

0.75

 LAST

0.71

hes

0.70

Activations Density 0.513%

words related to actions or accomplishments

past tense verbs associated with completed actions

No Comments

No Known Activations