INDEX

Explanations

words related to stopping or ceasing an action

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "halt" and its variations 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 @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Sov

-0.90

rics

-0.79

 Dynamics

-0.76

ãĤĮ

-0.72

çīĪ

-0.72

ophers

-0.69

iosyncr

-0.68

ramid

-0.68

aldo

-0.66

RAM

-0.65

POSITIVE LOGITS

ŃĶ

0.86

steen

0.83

 halt

0.82

ignt

0.79

 seiz

0.74

 derail

0.73

 halted

0.70

down

0.69

downs

0.67

 abruptly

0.66

Activations Density 0.018%

words related to stopping or ceasing an action

occurrences of the word "halt" and its variations in various contexts

No Comments

No Known Activations