INDEX

Explanations

past participles or verbs that indicate 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_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.10.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Wik

-0.74

RET

-0.73

yrinth

-0.67

USS

-0.65

sit

-0.60

 Seraph

-0.60

retch

-0.59

 Vlad

-0.59

 Brah

-0.58

uminati

-0.57

POSITIVE LOGITS

 during

0.77

during

0.75

 exceeds

0.73

per

0.72

aneous

0.67

gling

0.67

 outdoors

0.66

 indoors

0.65

itter

0.65

immer

0.65

Activations Density 0.653%

past participles or verbs that indicate completed actions

No Comments

No Known Activations